The effect of uncertainty on estimates of hurricane surge hazards

It is shown here that uncertainty can significantly affect estimated surge levels over a wide range of annual exceedance probabilities (AEPs). For AEPs in the range of 1 × 10<Superscript>−2</Superscript>–5 × 10<Superscript>−2</Superscript> in the New Orleans area, estimated surge values with and without consideration of uncertainty differ by about 0.5–1.0 m. Similarly, suppression of natural variability, such as using a single value for Mississippi River discharge in surge simulations, rather than allowing the discharge to vary probabilistically, is shown to produce deviations up to 1 m for the 1 × 10<Superscript>−2</Superscript> AEP in locations within the mainline river levees in this area. It is also shown that uncertainty can play a critical role in the analysis of very low probability events in the AEP range 1 × 10<Superscript>−4</Superscript>–1 × 10<Superscript>−6</Superscript>. Such events are typically used in designs of structures with major societal impacts. It is shown here that, for this range of AEPs along the west coast of Florida, the neglect of uncertainty can under-predict design surge levels by about 20 % compared to estimated surge levels that include uncertainty. Copyright Springer Science+Business Media B.V. 2013